【Press Release】Recursive Launches Weather Data Generation AI “Borealis”

This is an implied translation of a press release issued in Japan on March 23, 2023.

Even at locations far from the default weather observation points AI model developed that enables simulations with high accuracy of 100m mesh or better

Recursive has developed Borealis, which analyzes past weather data and generates weather information such as precipitation, temperature, and solar radiation at any location in Japan, and has begun providing this information to companies and local governments.

Borealis was developed based on various satellite data, AMeDAS and other meteorological data, and topography. The system analyzes information from weather observation points over the past 20 years, simulates weather information for any given point, and generates it using AI. Since the number of weather observation points in Japan is limited to a certain number of locations, it is not easy to obtain detailed weather data for points far from the observation points, and it has been common to spend manpower, time, and money to conduct surveys. Borealis supports quick decision-making by companies and local governments.

For example, Borealis can be used to select sites for small-scale solar power plants and wind power generation, which are expected to be in demand in Japan in the future. Borealis can provide highly accurate information on the amount of sunlight, wind direction, and rainfall, making it possible to simulate the amount of electricity generated in a shorter time. Borealis can also be used for a wide range of other applications, such as hazard mapping for natural disasters, farmland planning for optimal agricultural production, and development of regional cities.

Recursive is developing AI models for a sustainable society. Through Borealis, we will contribute to society in multiple ways, including the promotion of clean energy, disaster countermeasures, and regional development.

■Technical Structure